A Study on Whether Census Adjustment if Worthwhile

Written by:
RR86-17

Introduction

Suppose that we have a set of N disjoint blocks, indexed i=1, ..., N. One might think of these blocks as being the 50 states plus DC, but they might also correspond to smaller areas. For block i we let ti denote true population, and yi denote population as officially obtained from the Census. Let T = ti and Y = yi . We assume knowledge of yi (and thus Y) and T, but not of ti. Presumably the knowledge of T would come from an external source such as demographic methods.

Related Information


Page Last Revised - October 28, 2021